Usage of Ai in reviews

[PStonix]

Once again big thanks for the work You put into this. I hopoe it will serve someone more than only me.

This three way comparison really shows that DeepL is superior when it comes to translation. The best quality would be propably achived by asking LLM for cerection in Your native language verifing it and then using dedicated tool for translation (DeepL). Btw text translated by DeepL shows writen 100% by human on test in this case.

But that's a lot of additonal work...

I don't know if I'm gonna be using ai for correction but i'm sure i will switch to deepl for translation now...

 

[whowhawhy]

You have used "redact" and "redacted" --- they don't mean what I think you presume they mean.

ahaha - I would have fallen for the same trap, so now first thought "oh, maybe learned yet another way to detect German speakers", but as PStonix using Polish, that seems to be more of a multi-language different use of redact/or/ion (resp. Redakt/eur/ion).

this seems to be on the fence between

You must be registered to see the links

and the shared meaning also existing but the difference being in how likely it is to be assumed over other meanings in general or in specific contexts (see e.g. abstinence?).

Spoiler

You don't have permission to view the spoiler content. Log in or register now.

---

My initial gemini prompt

is it based on your, or others', experience in how to communicate with the ai to get the intended results?
you seem to be using a lot of calques from english created for the sole purpose of this review to imply (to a human, at least) what english words or phrases should be in the final result/translation, and obfuscate it by declining them according to polish grammar.

spellcheck your initial prompt too because there seem to be typos and grammar errors (inventing declension for the calques when the calques happen to already exist in your language and use a different one?) mixed in - you can use ai for this too (have it highlight or list the suggested corrections for you so you can confirm their validity yourself).

 

[PStonix]

...

is it based on your, or others', experience in how to communicate with the ai to get the intended results?
you seem to be using a lot of calques from english created for the sole purpose of this review to imply (to a human, at least) what english words or phrases should be in the final result/translation, and obfuscate it by declining them according to polish grammar.

spellcheck your initial prompt too because there seem to be typos and grammar errors (inventing declension for the calques when the calques happen to already exist in your language and use a different one?) mixed in - you can use ai for this too (have it highlight or list the suggested corrections for you so you can confirm their validity yourself).

This is based on my own experience in communication with the ai. I am using ai on daily basis. Typos don't really make a lot of impact on LLM responses, and yes I'm dyslexic and I've been too lazy to eliminate them, but i don't think this is too relevant in this context. I've wrote initial prompt before i had started this thread and i didn't exactly give a fuck about that at the time of writing it i am gonna look into some plugin to help me with that thank You for advice.

And about calques - again it doesnt make that big of an impact on LLM output and sometimes i feel like english word is a tiny bit more precise than the polish one so I use english onein place i don't normally write like that only to AI . Plus i tend to think partially in english when i plan for something to end up in english... It is complicated ok.